@palamedes/runtime

Exports

• getI18n<T>()
• setClientI18n(i18n)
• subscribeClientI18n(listener)
• setServerI18nGetter(getter)
• resetI18nRuntime()
• I18nInstance

The server subpath @palamedes/runtime/server exports:

• createServerI18nScope<T>()
• ServerI18nScope

Client Runtime

import { createI18n } from "@palamedes/core"
import { setClientI18n } from "@palamedes/runtime"

const i18n = createI18n()
setClientI18n(i18n)

Call setClientI18n() before translated client UI renders.

getI18n<T>()

Returns the active runtime instance. Transformed macro code calls this automatically.

On the client, getI18n() reads the instance registered with setClientI18n(). On the server, it reads the active server getter.

getI18n() throws a descriptive error when no active client instance or server getter result is available. Initialize the runtime before translated code runs.

subscribeClientI18n(listener)

Registers a listener invoked on every setClientI18n() call, including re-activation of the same instance in place. Returns an unsubscribe function.

This is the bridge framework bindings use to connect the framework-agnostic client runtime to their own reactivity system. @palamedes/solid/runtime, for example, feeds it into a Solid signal so translated output re-renders on a live locale switch. Application code rarely calls this directly.

const unsubscribe = subscribeClientI18n((i18n) => {
  // react to the newly activated client instance
})

Server Runtime

For request-local server rendering, prefer @palamedes/runtime/server:

import { createServerI18nScope } from "@palamedes/runtime/server"
import type { PalamedesI18n } from "@palamedes/core"

export const serverI18n = createServerI18nScope<PalamedesI18n>()

serverI18n.activate(i18n)

createServerI18nScope() uses Node AsyncLocalStorage, so keep it out of client bundles and Edge-only runtime paths.

Test Reset

resetI18nRuntime() clears the active client and server runtime state. It is intended for tests.